Use when you have a written implementation plan to execute in a separate session with review checkpoints
npx skills add niracler/skill --skill "anki-card-generator"
Install specific skill from multi-skill repository
# Description
生成 Anki 闪卡,输出 simple-anki-sync 兼容格式。Triggers on "anki", "flashcard", "memorize", "study cards". 当用户说「生成 Anki 卡片」「做卡片」「帮我记忆」「做闪卡」「学习卡片」「背诵卡」「记忆这个」时触发。
# SKILL.md
name: anki-card-generator
description: 生成 Anki 闪卡,输出 simple-anki-sync 兼容格式。Triggers on "anki", "flashcard", "memorize", "study cards". 当用户说「生成 Anki 卡片」「做卡片」「帮我记忆」「做闪卡」「学习卡片」「背诵卡」「记忆这个」时触发。
Anki Card Generator
Generate high-quality Anki cards based on LessWrong best practices and simple-anki-sync format.
Pre-Generation Clarification
Before generating cards, clarify with user if unclear:
- Scope: "Which aspects to focus on?" (for broad topics)
- Depth: "Basic concepts only, or detailed Level 2/3 cards?"
- Quantity: "How many cards? (Recommend 5-10 for core concepts)"
- Context: "Any specific exam or application scenario?"
Proceed only after understanding requirements.
Output Format
Use simple-anki-sync format:
#anki/[domain]/[topic]
| [Question] |
| ---------- |
| [Core answer]<br><br><small>💡 [Supplementary info]</small> |
Format Options
Option A (Recommended): HTML tags
| 唐朝建立时间 |
| ---------- |
| 618年,李渊建立<br><br><small>💡 隋末农民起义后起兵</small> |
Option B: Separator
| 唐朝建立时间 |
| ---------- |
| 618年,李渊建立 ——— 💡 隋末农民起义后起兵 |
Option C: Parentheses
| 唐朝建立时间 |
| ---------- |
| 618年,李渊建立(趣闻:其子李世民功劳最大) |
Atomization Rules
Word Limits
- English: Max 9 words, absolute limit 18 words
- Chinese: Recommended 15-20 characters, absolute limit 30-35 characters
- Max items: 3 bullet points per card
Core Principle
If a card can be split into two shorter cards, split it.
Question Design
Standardized Templates
- Time: "X 时间" (not "X发生于何时?")
- Definition: "X 定义" (not "什么是X?")
- Person: "who X" (not "谁做了X?")
- Pros/Cons: "X 利弊" (not "X的优势是什么?")
Key Rules
- Match real-world recall scenarios
- Use plain, unremarkable wording
- Avoid words in question that appear in answer
- Keep all critical info in answer, not question
Answer Construction
Core Answer
- Strictly follow word limits
- Answer should be meaningful without the question
- All key information in answer
Supplementary Info (Optional)
Format: <br><br><small>💡 content</small>
Emoji Guide:
- 💡 Fun fact / trivia
- 📝 Note / explanation
- 🔗 Related concept
- ⚡ Tip / key point
- 📊 Data / statistics
- 📅 Date / timeline
Keep supplementary info to 10-20 characters.
Handle System
Use > to reference related cards:
| 牛顿贡献 |
| ------- |
| >运动定律 >万有引力 >微积分发展 |
Detail Levels
- Level 1: Basic concept (core answer)
- Level 2: Detailed info (supplementary section)
- Level 3: Advanced details (create separate cards)
Tag Naming
Use English tags: #anki/[domain]/[topic]
Common domains: history, programming, language, science, mathematics, psychology, economics, philosophy, medicine, art
Quality Checklist
Per Card
- [ ] Core answer within word limit?
- [ ] Correct supplementary format?
- [ ] Can it be further split?
- [ ] Question matches real recall scenario?
- [ ] No memory shortcuts?
- [ ] All key info in answer?
Card Set
- [ ] Appropriate cross-references?
- [ ] Proper detail levels?
- [ ] No redundancy?
Domain Examples
See references/examples.md for detailed examples:
- History (ancient China)
- Programming (Python)
- Language learning (English)
- Academic concepts (psychology)
- Complex topics (quantum mechanics)
- Error corrections
- Advanced techniques (reversible cards, redundancy design)
Workflow
- Receive input: Knowledge points, wiki links, study materials
- Clarify: Ask questions if uncertain
- Generate: Follow atomization and best practices
- Output: simple-anki-sync markdown format
Core Philosophy: Prioritize sustainability and real recall scenarios over comprehensive coverage. Focus on preventing cognitive and motivational barriers.
When NOT to Use
❌ 不适合卡片化的内容:
- 需要视觉理解的(图表、流程图)
- 超过 50 行的代码块
- 高度情境依赖的知识
✅ 替代方案:
- 视觉内容 → 配图索引卡片
- 大代码块 → 代码 Snippet 库
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.